Methodology
How the calculation works and how we test ideas. These rules are common to all markets.
The Quasimodo (QM) pattern: left shoulder, structural low, a head that takes liquidity, and a structure break. The scenario: entry at the edge of the QML zone, stop beyond the head, target 2R. “Idea quality” is the number of confirmations minus twice the number of contradictions. We do not invent success probabilities: we only show frequencies from our own history, with the sample size.
- If one candle touches both the stop and the target, the stop is counted first.
- A target hit on the touch candle itself does not count; a trade is held for at most 300 bars.
- The default crypto fee is 0.08% per round trip, subtracted as “fee / stop size” in R.
- No knowledge from the future: every object has the time at which it became known.
- Rules are fixed in advance; thresholds are chosen on the first part of the time range (train) and evaluated on the last part (test).
- The number of variants tried, fees and survivorship of coins are taken into account.
- If an effect does not hold on the test part, it is treated as overfitting.
Over a year on 30 coins (about 7,650 trades) the pattern gives roughly +0.05R per trade before fees and roughly −0.11R after them. Filters by structure and by order flow (delta, CVD, absorption) did not hold up on the test part of the data. QM is a context tool, not a strategy. All checks: Crypto → Research.